# Genie 4 simulates worlds you can steer in real time By Steven Van ยท 2026-10-27 Google DeepMind's new world model is a research preview for trusted testers, with a waitlist before it reaches Project Genie and the Gemini API. Google DeepMind published Genie 4 on October 27, 2026, and called it the lab's most realistic and interactive world model yet. You explore a generated environment in real time: the scene responds, it makes sound, and the people in it move, react, and speak. The model is headed for [Project Genie](). [View media]() Google DeepMind's Genie 4 announcement film. Music, ambient noise, and dialogue are generated with the picture. Gemini Omni anchors how a simulation looks and behaves, and other Gemini models keep it coherent. You move with keyboard controls and change the world in natural language while it runs, including the weather, the camera, or a new person or object. DeepMind also describes stronger physics, and an experiment where Gemini 4 Argon writes a robot plan that Genie 4 simulates as it renders. That experiment is research, not a finished robot product. Genie 4 is a research preview for trusted testers. DeepMind says it comes next to Google AI Ultra subscribers in Project Genie and to early access partners of the Gemini API and Google AI Studio. A waitlist is open. The post does not give a date or a price, and it does not say Ultra access is available today. Details can drift in a longer session, real places are not simulated perfectly, a session lasts a few minutes, and real-time audio is still being improved. Video and audio carry a SynthID watermark. Sources [DeepMind's Genie 4 announcement](), [the Genie 4 model page](), and [DeepMind's post on X](). [Read Genie 4 simulates worlds you can steer in real time on Creators Toolbox]() --- Canonical source: https://creatorstoolbox.com/blog/genie-4-world-model